On Sun, Apr 4, 2010 at 12:36 PM, Stefan Küng <tortoisesvn_at_gmail.com> wrote:
> On 04.04.2010 19:28, Steve Borho wrote:
>>
>> Hello,
>>
>> The TortoiseHg project has been using your TortoisePlink application
>> as our default ssh transport for a couple of years. A while ago some
>> performance problems [1] were raised with it and just recently I had
>> the opportunity to investigate in more detail. I compared a clone[2]
>> operation using several tools and configurations, and this is what I
>> found (apologies for ascii table art):
>>
>> openssh -C: 0:30
>> openssh: 1:10
>> Plink: 1:20
>> paramiko: 2:00
>> TortoisePlink: 4:00
>>
>> I built Plink and TortoisePlink myself to get the tip version of each.
>> The clone time for both Plink and TortoisePlink was unchanged when
>> using the -C compression option. I suspect the compression was only
>> being used for upstream direction, but that's speculation on my part.
>>
>> The executive summary is there seems to be a lot to gain from pulling
>> the tip of Plink into TortoisePlink.
>
> I'm reluctant to update plink. They have a time machine on their server so
> you're never sure if you really get the latest release.
> See this thread for details:
> http://groups.google.com/group/tortoisesvn/browse_thread/thread/f2a890f517b835bc/b19072c65874f1bd?lnk=gst&q=0.60#b19072c65874f1bd
>
> That said:
> if you have a patch against plink 0.60 (if it really *is* 0.60 and not a
> beta), please send it here.
Youch. I admit I do not have access to their time machine either.
It is very tempting to spend a weekend adding a simple windows UI to
openssh and be done with this.
--
Steve Borho
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2483368
To unsubscribe from this discussion, e-mail: [dev-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2010-04-05 07:53:03 CEST